Hacker News Reader

Well for example you can take several of the clocks and compare them.

Would that require the assumption that any inaccuracies affecting multiple clocks are uncorrelated with each other?